Sponsored by: Corsair’s TradeCraft (now in Early Access)
The Practical Explainer — facts
Simplifying the landscape of analytics — one concept at a time
It is amazing how many words in analytics can mean the same damn thing. Case in point - facts, metrics, and measures are essentially interchangeable. At the most basic level, they are all just another name for numbers. In the database universe, number is a format but facts are a data element. Take a fact, put it in a spreadsheet table, suddenly you have a measure… or is it a metric. Add it to a pivot table — now it is a value. Is it all just semantics… mostly yes. Let’s just start with that.
Breaking It Down Further:
When you see a number referenced with the terminology fact, you are probably deep in the world of referential databases, OLAP cubes, or star schemas. We will need other articles to delve through all of that, but it is worth noting that facts are often defined in direct contrast to dimensions. Dimensions absolutely require another article entirely, but have their own roster of synonyms including segments, groups, attributes, and labels.
In practice:
Facts (or whatever you choose to call them) are best understood as the numbers, values, or measurements that simple mathematical calculations can be applied to. Said more plainly — facts can most often be summed, averaged, and counted. Dimensions are most often grouped, sorted, and separated in ranges. As is often the case — some data elements fit both definitions. If that last part is confusing — consider age. Age can be summed or averaged, but it can also be used as a dimension — typically grouping it in ranges. Typically, age would be stored separately for each purpose.
In this example, price and quantity would be described as measures or metrics (the more common spreadsheet terms). In this case, item is dimension.
If you are also noticing the alpha vs numeric nature of this distinction, you aren’t wrong. The distinction however is one of likelihood. While it is most often the case, there is no guarantee nor defined rule. Age is again a great example. Zip code perhaps a better one.
It is also fair to note that values (the pivot table variation) are most often associated with cells, whereas dimensions occupy columns and rows. But our simple example shows you how readily that is broken with the values appearing across the columns. But the rule does often apply.
Some final associations:
We have covered most of these already. This topic comes with so many associations, it is difficult to declare any finality. But there are still a few more worth mentioning — KPI or Key Performance Indicators are highly associated… perhaps too highly. You will similarly hear reference to Level I or II metrics along that path. This is honestly a bastardization of a dimension, but it will have to wait for our KPI article.
So just remember:
Facts, measures, metrics, and values are all just references to numbers in one context or another. They are intended to feed calculations and are therefore stored differently then their counterpart the dimension.
Thanks for reading. More Practical Explainer coming soon…
Want to train, learn, and practice calculating facts and measures? Checkout our sponsor.
